Traffic Sign Detection using CNN

This project uses a Convolutional Neural Network (CNN) to detect and classify traffic signs. It is designed as a prototype for car dash cams and currently supports processing pre-recorded video files. The output video highlights detected traffic signs and displays labels or warnings.

Features

• Traffic sign detection from video files.
• Saves the labeled output for further analysis.
• Modular scripts for training the model and running inference.

Planned Enhancements

• Real-time traffic sign detection using live video streams.
• Improved performance and detection accuracy.

Technologies Used

• Python: Core programming language.
• TensorFlow & Keras: For creating and training the CNN model.
• OpenCV: For handling video input, frame-by-frame detection, and saving annotated video output.
• Google Colab: Platform for training the CNN model with GPU acceleration.

Data Sources

• Traffic Sign Dataset: Downloaded from Kaggle.
• Input Video: Sample videos for testing obtained from Kaggle.